To what extent are health selection processes responsible for adult health disparities among different socioeconomic groups? Is it possible that at least part of the long-lasting presence'of quite vigorous adult health and mortality inequalities in the US and other developed countries owes a non trivial contribution to processes whereby individuals who experience early child health problems end up in lower ranking classes as well as in poorer health status? And if so, is it a large or a small part? This project attempts to answer these questions for individuals who belong to the high school class of 1957 in the Wisconsin Longitudinal Study (WLS). In order to do so, we undertake three different tasks. The first task is to empirically estimate relatively 'pure'effects of education and wealth on adult health status and mortality, that is, purged from contamination due to unobservables. The second task is to estimate alternative models for the production of adult health and mortality that take account of early conditions and upbringing. The estimates of these alternative models can then be deployed (via Monte Carlo simulation) to classify the estimated effects of education and wealth into two categories: the first includes only direct effects, that is those associated with characteristics inherent to wealth and education;the second class is comprised of 'indirect effects', namely, those resulting from mechanisms that link health status and mortality as well as wealth and education to processes that occur early in the life of the individual. The third task we propose to accomplish is to evaluate if and to what an extent the results we obtain are peculiar to the WLS (given its unique configuration) or rather, apply more generally to the US population. For this purpose we use the cohorts followed in the Health and RetirementSurvey.